基于α相同度相似关系的rough集模型  

Rough set models based on α-identical degree similarity relation

作　　者：周辉[1] 王黔英[1] 费颖[1] 袁芳[1] 

机构地区：[1]南昌大学管理科学与工程系,江西南昌330047

出　　处：《计算机应用》2006年第3期666-667,共2页journal of Computer Applications

摘　　要：Rough集理论是一种处理不完备信息系统的数学工具,但是Pawlak的经典rough集理论似乎是不可行的,因为它要求论域中数据间有很强的等价关系。在产生基本集(相似类)时,一般相似关系的分类误差较大,集对分析会把两个对立度不为0的个体划分在一起。汲取两者的优点,给出相同度的概念,只有满足一般相似关系并且相同度大于或等于阈值α的两个对象才能划分在一个基本集中,在此基础上建立基于α相同度相似关系的rough集模型。通过实例验证效果要比基于集对分析或者一般相似关系的模型更好。Rough set theory is a mathematical tool to deal with incomplete information systems, but Pawlak＇s classic rough set model is unfeasible, because it requires strong equivalence relations among the datum of universe. When forming element sets （ similar class）, common similarity relation may have more error in classifying, set pair analysis may classify two objects that contrary degree between them is not ‘O＇ into a class. To solve above problem, concept of identical degree was put forward. Only if two objects satisfied common similarity relation and their identical degree exceeded （ or equaled to） threshold, they would be classified into an element set. The rough set model based on α-identical degree similarity relation was built. Experiments show that it is better than the model based on set pair analysis or common similarity relation.

关 键 词：ROUGH集 α相同度相似关系 β变精度 不完备信息系统 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]